We aren't currently maintaining Metaforecast. We hope to do so again in the future.
OpenAI's best released model could be GPT-4, GPT-4o, or something else. It does not count as an OpenAI model unless it's made available to the public to try, and is known to be from OpenAI (e.g. the model can not be a secret, pseudonymous release). If arena.lmsys.org is not available at the time, the successor site or most similar leaderboard will be used.
Resolves yes if Claude 3.5 Opus is ranked above all OpenAI models 1 week after it is put on the leaderboard.
Update 2025-01-01 (PST) (AI summary of creator comment): - Models must be listed on lmarena to be counted.
Examples:
o1 pro does not count since it's not on the arena.
Regular o1 does count.
| Indicator | Value |
|---|---|
| Stars | ★★☆☆☆ |
| Platform | Manifold Markets |
| Forecasters | 49 |
| Volume | M5.4k |
OpenAI's best released model could be GPT-4, GPT-4o, or something else. It does not count as an OpenAI model unless it's made available to the public to try, and is known to be from OpenAI (e.g. the model can not be a secret, pseudonymous release)....